Empathy is a Must for Machine Learning SAP Mentor Call Recap
Title: Empathy is a Must for Machine Learning, February Community Call
Abstract: Artificial Intelligence, Machine learning and Predictive Analytics are at a perfect storm and many companies leverage these technologies to transform their organization. These technologies existed for a decade, but they evolved rapidly – machine learning today is not like machine learning of the past. In these session, we will look into these technologies, understand what SAP is to offer and how SAP customers are using these technologies.
Call Recording: Part 1, Part 2
Hosted by: Chris Kernaghan
Speakers: Chandran Saravana, Senior Director, Advanced Analytics SAP
Markus Noga, Innovation Network
Figure 1: Source: SAP
Emotions are a big business
Growing market and how enable empathy in machine learning
Figure 2: Source: SAP
Start with empathy for the end user
Figure 3: Source: SAP
What is empathy?
Figure 4: Source: SAP
What are the markers? Shown in Figure 4
How bring empathy to it – structured and unstructured data
Figure 5: Source: SAP
Why important now? Computing power
Genome analysis – took days to do
Figure 6: Source: SAP
Early stages for empathy for machine learning
Figure 7: Source: SAP
Use cases for machine learning – teaching, simple math lessons, can be done by a robot
Figure 8: Source: SAP
Goals of empathetic machine learn
Respond intelligently to emotions
Figure 9: Source: SAP
Universal facial expressions
7 primary emotions and other secondary emotions
Figure 10: Source: SAP
4 knowledge approaches are shown in Figure 10
Figure 11: Source: SAP
Algorithms used
Figure 12: Source: SAP
Requires speed and agility
Figure 13: Source: SAP
Definition of digitization is shown in Figure 13
Figure 14: Source: SAP
Digital framework, machine learning is a part of it
Figure 15: Source: SAP
SAP Clea – making enterprise applications intelligence
Figure 16: Source: SAP
Trying to automate knowledge work
Take load off customer service so humans can handle complex situations
Figure 17: Source: SAP
Use cases across SAP
Figure 18: Source: SAP
Use cases
Figure 19: Source: SAP
Roadmap is subject to change
Figure 20: Source: SAP
Apps are for the business users
API’s are for the developers
Training infrastructure is for data scientist
Questions
Q: Are there differences in training the empathic model versus normal statistical mode
A: Key element is data availability and images
Q: Taking open platforms into the context, how does sap ml pair with the data lakes and all platforms out there beyond vora?
A: Going natively on SAP Cloud Platform; will be natively connected on SAP Cloud Platform
Create machine learning platform partner program; increased openness
Q: Where is data held, privacy/data concerns..
A: data protection is serious; follow German privacy laws, which are strict; linked to SAP Cloud Platform roadmap
Q: How guard against biases? Creating new ones?
A: Discussing with SuccessFactors colleagues and Business Beyond Bias campaigns and machine learning can contribute
Try to find calibrated data
Q: Are these models pretrained?
Or does the customer have to do their own training?
A: Some things are pre-trained, others are not
Q: Relationship between Clea and Leonardo?
A: Clea will help all SAP apps, including IoT